A Dynamic Scheduling Workflow Algorithm Based On Critical Path

• Published in: Journal of physics. Conference series Vol. 1994; no. 1; pp. 12039 - 12047
• Main Authors: Tongyue, Zhu, Wenhui, Yang
• Format: Journal Article
• Language: English
• Published: Bristol IOP Publishing 01.08.2021
• Subjects: Algorithms; Cloud computing; Completion time; Critical path; Dijkstra's algorithm; Nodes; Optimization; Priority scheduling; Resource scheduling; Scheduling; Task scheduling; Workflow
• DOI: 10.1088/1742-6596/1994/1/012039
• ISSN: 1742-6588

Abstract With the explosion in the size of scientific workflows, local workflows can no longer satisfy existing computing needs, and cloud computing platforms have become the first choice for scientific workflows. Compared with traditional local computing, cloud computing platform not only needs to consider the loss of scheduling time and transmission time, but also involves the cost of cloud resources. Therefore, whether the scheduling scheme is reasonable becomes the decisive factor of workflow efficiency. Aiming at how to achieve efficient scheduling under cost constraints, a dynamic scheduling algorithm based on critical path is proposed. The algorithm uses the optimized Dijkstra algorithm to classify the task nodes, guarantees the completion time of key nodes in the scheduling process, adjusts the priority of nodes dynamically, and selects the best resources according to the loss weight. Experiments show that the optimization rate increases with the increase of the number of task nodes. The algorithm proposed in this paper is suitable for large-scale workflow operation and can effectively reduce the scheduling time.